LETTER TO THE EDITOR

HTL V -I and CTCL: The Link is Missing

To the Editor:

We read with great interest the paper of Khan et al (1996) describing the localization of human T-cell Iymphotropic virus type-I (HTL V -I) tax proviral sequ ences in skin biopsies of mycosis fungoides utilizing ill silll polymerase chain reaction (PCR) . In their stud y, 11 of 12 skin biopsies from patients with mycosis fungoides demonstrated positi ve tax signal localized over dermal and epider­mal infiltrating lymphocytes. T hey sta te " demonstration that the lymphocytes in the skin bi opsies of these patients harbor H TL V-I proviral sequences seems beyond cavil. " Yet, they fail to dem on­strate rhe specificity of the tax signal through the use of appropriate negative controls (a reaction without Taq polymerase or with an irrelevant primer pair or with a nonspeci.fi c probe). In som e cases, tax signal was detected over keratinocytes ill les ional skin; how­ever, a comparison with negative controls (normal o r unin volved skin) a nd positi ve controls (skin fi·om a seropositive HTLV-I­infected patient) is not provided . T he authors conclude that ill Silll PCR studies of HTLV-l tax proviral sequences can help establish the diagnosis of mycosis fungoides. We take exception to their conclusion because a critical evaluation of the published reports in the world's literature will support the conclusion that H TLV-I docs not playa role in the pathogenesis of cutaneous T -cell lymphom a (mycoses fun goides/Sezary syndrome)(CTCL).

In six series (Capesius et ai , 199 1; Lisby et ai, 1992; Bazarbachi ct ai, 1993; Boni CI ai , 1996; Li Cl ai, 1996; Wood ci ai, 1996) PCR amplification was utilized to detect HTL V -I proviral sequences in a total of 205 patients fi·om Euro pe and the United States. No

I evidence of HTLV-I proviral DNA was detected in any of the patient s screened (01205). In ano ther fo ur reports (D'lnc:ln et ai,

I 1992; Srivastava et al. 1992; C han el ai, 1993; Whittaker and Luzzatto, 1993), a total of176 patients from Europe and the United States were screened. I-ITLV-I proviral sequences were detect in 18 patients (10%). Two studies (Ghosh et ai, 1994; Pancake el ai, 1995) have screened a total of 75 patients from the United States and detected I-ITLV-I proviral DNA in 64 patients (85'X) . Overall , the detection rate is 18% (84/456).

To date, no definitiv e model of HTL V -I infection in CTCL has been a d vanced or confirmed . T runcated proviruses and low copy number (Hail et ai, 1991 ; Z ucker-Franklin el aI, 1991) have been advan ced as factors influencing detection rates of HTLV -I and the lack of antibody response seen in CTCL (Srivastava et al. 1992). If such mechanisms are associated with HTL V -I in fection in CTCL, then an entirely new and yet to be confirmed I-ITLV-I virology mus t b e operative in CTCL, because it is neither endemic for areas ofHTL V -I infection (Weinstock and Homl, 1988) nor does CTCL have the clinicopathologic features of endemic I-ITL V -I-associated adult T - cell leukemia/ lymphom a in the majo ri ty of cases (DiCaudo et ai, 1996).

The occasional detection of H TLV-I proviral DNA in CTCL patients m ost likely is a result of one or more facto rs: (i) inaccurate diagnosis of CTCL in some patients (i.e. , inclusion of one 0 1· more patients likely to have adult T -cell leukemia/ lympho ma based on their classic HTLV-I seropositi vity [e.g., Pancake 1ft ai, 1995)]; (ii) second ary infection related to CTCL immun osuppression (Heald c/

ai, 1994); (iii) ampl.ifica tion of endogeno us retrovirus-like genes (Mager and Freeman, 1987); (iv) trace con tamination of samples (Wood el ai, 1996). T he high detection rates reported in on ly two

studies requires careful interpretation (Ghosh e/ ai, 1994; Pancake et

ai, 1995). Fint, none of these studies demonstrated genomic integration of I-ITLV -I. precluding any definitive conclusions re­garding a potential mechanism or site of HTLV -I infection in CTCL. O ne of these groups (Ghosh e/ ai, 1994) has previously published similar PCR detection of HTLV-I DNA in multiple sclerosis (Reddy et aI, 1989), which subsequent studies have failed to confirm (Winkelman, 1993). The laboratOl), of Kllan et ai, although publishing PCR detection rates of 92% in tlIe peripheral blood (Pancake et ai, 1995), has never correlated or tested their PCR methods in skin. Instead, they introduce the use of i ll sitll PCR to detect I-ITL V -I DNA in CTCL without appropriate controls and without addressing their detection rate of whole-tissue PCR am­plification in their skin biopsies.

Based on the currently available m olecular and epidemiologic data, .it is clear that H TL V -I is nOt a primary etiologic agent in the pathogenesis of CTCL. We therefore feel that tile use of ill 'sit ll PC R detection of HTL V -T tax proviral seq uences as a diagnostic tool in CTCL is not va lid .

Stua rt R . Lessin , Alain I-I . Rook, Guoqing Li Departm ent of Dermatology

University of Pennsylvania & the Phil adelphia V A Medical Center

Philadelphia, PelIDsylvaniu

Gary S. Wood Department of Dermatology

Case Western Reserve University & the Cleveland V A Medical Center

Cleveland, O hio

REFERENCES

B:\z:lrbachi A. 5a:11 F. L:lfochc L. Flagcu l 13, PcricsJ, de The H: HTLV-r provirus and mycosis limgoidcs . Sricllce 259: 1470-147 1. 1993

Doni R~ Dnvis-D.\1lcshf.,r A. Burg G. Fuchs D. Wood G: N o detection of HTLV-I proviral DNA in Icsional skin bi opsies from Swiss anei German patients with cutallCOLIS T -cell lymphol1la. Bt) Oel"ll"101 136:282-28'1. 1996

Capcsi ll s C. Saal F. Macro E. B azarbach.i A, Lasncrct J. Laroche L. Gcssain A. I-Iojman F. Peries J: No evidence for HTLV- l infection in 24 cases of French and Portuguese mycosis fu ttgoidcs flnd Sczary syndrome (as seen in France). U:ll k'clllin 5:4 16-·11 9.1991

C h ilT1 \ ,(/, Hooper C . \Xfichcrt R. Benson J. VardimJI1 J. Hinrichs S. Weisenburger D : HTLV-I sequeuce in IYl1l l'hoprolifcrativc disorders. Dio_~11 M"I POII<0/2:192-199 , 1993

O' incan M. SOUlhtcyr:1I1d P. Bignoll Y. Dastugllc B. Cliludy A. Dcsgrangcs C: P .... ctrovirttl sequences rc larcd to HTLV -I endemic areas. ellr J DCfmnw/ 2:363-37 1. 1992

D iCa udo D. Perniciaro C. \'(forrell J. \ Vhi rc J. Cockc.rcil C: Clil1ic~1I and histologic spectrum of human T-ce ll lymphotropic virus type I-as ocinted lymphoma involving .. he skin.) Alii Arlld Oem",,,,1 34:69-76. 1996

Ghosh S. AbramsJ. TCrl.1I1 UII13 H. Vondcl'i1cid E. DeFreitas E: Human T-ccllleukcmia virus type I tax/ rex DNA and P.....NA in cutaneous T -cell lympho ll1:l. Blooil 84 :2663-2671. 1994

Hall WW. Lill C R . Schnccw ind O . Takahashi H . Kaplon MH. R o upe G. Vahlnc A: Deleted HTLV-I provinls in blood and c utaneous lesions o f patients with mycosis fungoide,. SciClI"" 253:3 17-320. 199 1

I-ic:dd P. Y,lIt S-L. Edelson It: Profound ddicicnt.'y in nortnal circ lll:lting T ce lls in crytllrodcrl11ic cuta ll COUS T-cell IYl llpholll a. A,r!J Df.,."WlcIJ 130: 198-203. 1994

Khan Z M . SC bCll ik M. Z uc ker-Franklin D: Loc;)liza tiol1 ofhulllan T-ccll lytnphorropic virus-I tax prov iral scquences in skin hiopsies of p:lticllts with m ycosis fl.lOgo idcs by itt sit" polyl11 c r~I Sc chain I'c:lctioll .J llllll'st D('rII /("ol 106:667-672. 1996

0022-202X/96/S10.50 • Copyright 1996 by T he Society for In vestigative J)crm:ltology. Inc.

783

784 LETTER TO THE ED ITOR

Li G. Vowels llR. Benoit BM. Rook AH. Lessin SR: Foilure to detect human T-lymphotrophic virus typc-I proviral DNA in ce ll lines and ti ssues fro III p:,ticnts with CUWIl COUS T-cell lymphoma. J lll llest D ermalof 'I U7:308-3 J 3, 1996

Lisby G. Reitz MS. Vejlsgoord GL: No detection of HTLV- I DNA in punch skin biopsies fro 111 patients wi th cutaneo lls T-cell Iympho m., by the PQ l ynlc r~l sc chain reaction. J IlI ltest D enllnlol 98:417-420. 1992

Mager DL, Freeman JD: Human endogenous rctroviruslikc genome with type C pol sequences and gag sequences related to human T-cell Iymphotropic viruses. J Viml 61 :4060-4066. '1987

Pancake BA, Zuker-Franklin D, COllt01V3S EE: The c utalleous T ccB lympho ma. mycosis fungoidcs. is a human T celll Yl11 photropic virus-associated disease. ) e li" Itlllesl 95:547-554.1995

Reddy E, Sandberg-Wolheim M, Mettus R., Roy P, DeFreitas E, Koprowski H : Amplification and molecu l<1r c10nig ofJ-ITLV-1 sequences &0111 DNA of multiple sclerosis patients. ScicI1cC 243:529-532, '1989

Srivastava SE I, Oanki K, Perl A: Human T- cell leukemia virus type 1 or a re lated retrovirus in p;ltients with mycosis fUl1goidcs/Sczary syndrome and Kaposi's "rcolno. C allw' Res 52:4391-4395. 1992

Weinstock MA. I-I orm JW: Mycosis fungoidcs in lhe United States . Increas ill g: in cidence ond descriptive epidcl1lio lgy.JAMA 260:306-308, 1.988

Whittaker SJ. Luzzatto .L : HTLV-I provirus and mycosis fungoides. SciC:IICC 259: 1.470 . 1993

Winkelmon j: I-ITLV-1 ond mu ltiple sclerosis, J Lab Ciill Mcd 122:230-23'1. 1993 Wood G. Salvek" r A. Schaffer j . C rooks C . Henghold W , Fivenson D. Kill! Y-I-I.

Smaller B: Evidence aga inst a role for HTLV-I in the pathogenesis of Arncrit:all cu taneous T-cell Iymphomo , J lI111esl DWIla/o/l 07:30'1-307. 1996

Zucker-Frankli n D. CO UW V:'IS EE, Rush MG, Zouzias DC: Detection of human T - Iymphotropie virus- likc particlcs in culturcs of peripheral blood 1),lllphocytes of patients w ith mycosis fungoides. Pmc NIlII Acari Sci USA 88:7630-7634. 1991

Reply: Medical science does not advance by turning the clock backward. N ew methods, more sensitive methods, or, as in this case, a change in methods has yielded results which have been sought by a large number of investigators, many of whom have been cited in the letter submitted by Drs. Lessin, Rook, Li, and Wood. Negative results can, of course, not prove anything. In addition, nobody would claim, at the present time, that HTLV -I tax is causally related to mycosis fungoides. However, that such proviral se­quences are harbored in the lymphocytes of most of these patients (whether in th eir blood lymphocytes or those infiltrating the skin) is beyol/d ca lli/ .

Tt is more constructive to learn why so many in vestigators have failed to detect these sequences and why so few patients with bona fide CTCL have antibodies to the structural proteins of the virus.

We have meticulously carried out such an analysis on 10 specimens comparing our methods with those used by others, including those used in the papers cited by Drs . Lessin el al (Pancake and Zucker-Franklin, in press). The major differences are as follows: (i) the state and condition of the cells used , i. e ., freshly isolated lIerSIIS cultured or cells fixed in various ways. The la tter applies particul arly to skin biopsies that are almost always negative unless the specimens are fresh ly obtained or frozen; (ii) preparation of specim ens for PCR, i.e ., whole celllysates verslls DNA extracts; (iii) the HTLV sequen ce targeted, i.e., tax lIerSIIS other regions of the HTLV-proviral genome; (iv) whether or not Southern hybl;d­ization is used in combination with PCR. The use of DNA extracts rather than whole-cell Iysates may be the most important short­coming in the m ethods used by less successfill authors. A paper detailing these various methods is in press (Pancake and Zucker­Franklin, in press).

THE J OURNAL OF INVESTIGATIVE DERMATOLOGY

Since the vast majority of patients with CTCL have HTLV-I proviral sequences in their pel-ipheral blood lymphocytes (Pancake et nl, 1995), it should not be surprising to find these also in lymphocytes infiltrating the skin. We have, of course, done PCR/Southern ana lyses on skin biopsies (e.g. , Zucker-Franklin el

aI, 1994), but this exercise did not identify the specific cells, harboring the viI·al sequences, which was done in the paper ullder discussion.

The concern over the finding of BTL V-I tax sequences in the lymphocytes of patients with m ycosis fungoides may be alleviated by our observation that all patients who have demonstrable tax sequences also have tax mRNA (Pancake el aI, 1995) and the majority have antibodies to tax without having antibodies to the structural proteins of the virus (Pancake el nl, 1996; Pancake el nl, in press).

Indeed , it is unfortunate that the correspondents raise questions that have long been settled by others. For instance, they question whether we are detecting endogenous retroviral sequences when it is generally recognized that tax, the putative HTL V -transforming sequence, does not share homology with any human DNA se­quences. They also question whether the diagnosis of CTCL was accurate when the dermatology department at N ew York Univer­sity is one of the largest referral centers for this disease in the country . (By now we have studied upward of 150 mycosis fun­goides patients -mostly Caucasian Americans who are positive for HTLV-I tax sequences). Mycosis fungoides does not seem to be an endemic disease, but the same pathogen , HTLV-I, has been shown to be associated with diseases as divergent as ATL and TSP/HAM. Let us now move forward!

Dorothea Zucker-Franklin Department of Medicine

New York University School of Medicine New York, N ew York

REFERENCES

Pancake EA. Wasse r EI-I . Zucker-Franklin D: Patients Wilh cutaneouS T cell l)'tll­phorna (CTCL) tbought to be scrollcgntivc for I-ITLV-I. frcqucl1r1y have antibodies to HT LV tax. 810",/86:272. '1995

Pancake BA , WasscfEH. Zucker-Fran klin 0: Demonstration of anti bodies to HTLV-I tax in patients widl thc cutaneOli S T edt lympho ma , rn ycosis fungoidcs , who arc seronegarive for ill1obodies to rhe structural proteins of the virus. Blootl1996 in press

Pan cake OA. Z uckcr- Fmnklin D : The diffic ul ty of detecting HTLV-J proviral se­quences in paticllts with mycosis fungoidcs. J ACI}lIir Dejic S }llUlr HI/Ill Rern",jn?l, t 996 in press

Pane'lkc BA, Zucker-Franklin 0 , Colltav<ls E: The cutaneOllS T cc U lymphol113. mycosis fungoidcs is an HTLV-associatcd disc;\sc. A study of 50 patien(s. j Clilt 1I",,;sl 95:545-554.1995

Pancake UA t Zucker-Franklin 0, MOirmor M. Legler PM: Determination of the true prcv:llencc of in fccti 0 11 wi th the hum an T ceU Iymplto tropie viruses (HTLV-illl) may requirc a t:ornhination of biomolecular and scrologic all alyses . J Jill /cs t Mcd 44:250. 1996

Zuckcr-Fn:lIlkLin D, Pallcnkc BA. Fricdrn~II1-Kicn AE: C utancous disease resembling myt:os is fungQidcs in HI V-infected patients whose skin and blood cells aI 0

harbor proviral HTLV type J. A IDS Res HI/Ill l{ctn",j,."scs 9: '1173-1177. \99-'

J